Proofpoint: Generative AI Is Wiping Out the Behavioral Tells Email Filters Rely On

Generative AI has changed what a malicious email looks like. Messages that once betrayed themselves through clumsy wording and reused templates now arrive fluent and tailored to each recipient, and the behavioral signals email filters have leaned on for years are fading as a result, Proofpoint said at its Protect 2026 conference.

One example: a single threat actor contacted more than 30 real estate agents at the same company, each from a different account, about a different property, in different words, before closing every thread with the same request to join a video meeting. Because each sender was making first contact and every message differed, there was no history or repeated content for behavioral filters to match.

Federal data points the same way. The FBI's Internet Crime Complaint Center counted more than 22,000 complaints involving AI in 2025, with nearly $900 million in associated losses, plus 24,768 business email compromise complaints and about $3.05 billion in reported losses, according to its 2025 Internet Crime Report.

Proofpoint argues the analysis has to shift from asking whether a message looks unusual to asking what it is trying to do. It positions intent analysis as a third detection layer alongside traditional rules and signatures and modern behavioral baselines, targeting AI-crafted messages that pass both. The company announced its Nexus Intent-Based Detection Model in September as part of its Agentic Collaboration Security system.

Its Threat Investigation Agent is designed to run one investigation across email, identity, messaging, browser and supplier activity. In one example, an RFP-themed credential-phishing lure led to the takeover of a finance user's account; the agent contained the incident four minutes after the trigger, while four higher-consequence actions, including a credential reset, were held for human approval.